Update and clean-up compsets for use in test suites #4488

wlin7 · 2021-08-24T20:40:24Z

Remove obsolete eam and allactive compsets and the associated eam
use_case files, and update the compsets used in nightly test suites.

All compsets referring to AV1C or v0atm are removed. F20TRC5-CMIP6
and F2010SC5-CMIP6 are no more needed (previously retained for ongoing
testings). Remaining use case files referring to av1c are removed, including
those used by MMF tests. Also Remove obsolete use_case files referring to
cam3 and cam4 .
This PR also makes all standard F-cases use prescribed data ice mode of MPASSI.
This includes F1850, F1950, F2010, F20TR. The corresponding compsets that
use CICE are renamed as F1850-CICE, F1950-CICE, F2010-CICE, and
F20TR-CICE, respecticely. The F####-CICE compsets set to go
with SROF while the MPASSI counterparts with MOSART as river component.

The following compsets are replaced in nightly test suites.

. FC5AV1C-L with F2010
. A_WCYCL1850 and A_WCYCL1850S_CMIP6 with WCYCL1850
. FC5AV1C-H01A with F2010 as there is no v2 compset just for hires
. A_WCYCL1950S_CMIP6_HR and A_WCYCL2000_H01AS with WCYCL1950 as a placeholder

With compset changes and the use of MPASSI data ice mode, F2010 tests using ne4_ne4
change to use ne4_oQU240, and ne4pg2_ne4pg2 to ne4pg2_oQU480;
F2010 tests with ERS_Ln# and ERP_Ln# change to use ERS_Ld3.

[non-BFB] Virtually for all tests involving the compsets above. Tests using the old compsets would appear missing
while new tests would appear correspondingly.

Note: The PR message is updated (new text in the 2nd and 3rd paragraph)

Remove obsolete eam and allactive compsets and the associated eam use_case files, and update the compsets used in nightly test suites. All compsets referring to AV1C or v0atm are removed. F20TRC5-CMIP6 and F2010SC5-CMIP6 are no more needed (previously retained for ongoing testings). Temporarily leave F2010SC5-CMIP6-MPASSI which is to be replaced with F2010-MPASSI via #4483. Remainging use case files referring to av1c are removed, except for two used by MMF tests. The following compsets are replaced in nightly test suites. . FC5AV1C-L with F2010 . A_WCYCL1850 and A_WCYCL1850S_CMIP6 with WCYCL1850 . FC5AV1C-H01A with F2010 as there is no v2 compset just for hires . A_WCYCL1950S_CMIP6_HR and A_WCYCL2000_H01AS with WCYCL1950 as a placeholder [non-BFB] Virtually for all tests. Tests using the old compsets would appear missing while new tests would appear correspondingly.

rljacob · 2021-08-24T20:46:40Z

cime_config/allactive/config_compsets.xml

-  <alias>A_WCYCL1850-DIB-ISMF</alias>
-  <lname>1850_EAM%AV1C-L_ELM%SPBC_MPASSI%DIB_MPASO%IBISMF_MOSART_SGLC_SWAV</lname>
-</compset>
-
 <compset>
  <alias>A_WCYCL1850-DIB_CMIP6</alias>


Can we rename this one to WCYCL1850-DIB ?

@jonbob , can we remove this one? Look like CRYO1850 is the new equivalent?
Similarly, do we need to do something about A_WCYCL1850-DIB_CMIP6?

@wlin7 - which one do you want to remove?

I'm asking the cryo people about the older compsets -- the A_WCYCL1850-DIB-ISMF_CMIP6 is almost exactly the same as the new CRYO1850 (only different in specifying spunup ocn/ice IC's)

The cryo people would like to keep the A_WCYCL1850-DIB-ISMF_CMIP6, though it's also OK to drop the "A_" prefix

A_WCYCL1850-DIB_CMIP6 and A_WCYCL1850-DIB_CMIP6 are now removed.

rljacob · 2021-08-24T20:47:28Z

cime_config/allactive/config_compsets.xml

- <lname>2000SOI_EAM_ELM%SP_MPASSI_MPASO_MOSART_SGLC_SWAVi</lname>
-</compset>
-
-
 <!-- E3SM BG compsets -->

 <compset>
  <alias>A_BG1850CN</alias>
  <lname>1850_EAM_ELM%CN_MPASSI_MPASO_MOSART_MALI%SIA_SWAV</lname>


Lets remove the "A_" from this one.

rljacob · 2021-08-24T20:54:56Z

What we should do to verify bfb after the name changes is run cprnc by hand for a few cases on Chrysalis.

jonbob · 2021-08-24T20:57:35Z

cime_config/allactive/config_compsets.xml

-  <alias>A_BGWCYCL1850</alias>
-  <lname>1850_EAM%AV1C-L_ELM%SPBC_MPASSI_MPASO_MOSART_MALI%STATIC_SWAV</lname>
-</compset>
-


We may want to replace this one with something else

How about making it BGWCYCL1850_CMIP6 and changing "EAM%AV1C" to "EAM%CMIP6"?

A_BGWCYCL1850 has been removed. BGC compsets for v2 are yet to be finalized.

wlin7 · 2021-08-24T21:02:28Z

What we should do to verify bfb after the name changes is run cprnc by hand for a few cases on Chrysalis.

Good idea. How long the baselines for the would-be missing tests be saved? I would certainly do cprnc before merging to master and requests for new baselines.

Edited: Only those currently using A_WCYCL1850S_CMIP6 can be compared (and BFB expected). Not even F2010SC5-CMIP6 which has a couple parameters (via 2010_cam5_CMIP6) that would override the default used for F2010.

brhillman · 2021-08-26T21:49:17Z

@wlin7 this is great, thanks for doing this! We can fix the MMF compsets to also use the 2010 forcings and rename the use case files. Would you like me to push those to this PR, or do that in a future PR? @whannah1 may have thoughts on that too but I believe he is off today and tomorrow.

wlin7 · 2021-08-27T00:08:38Z

Thanks for the review, @brhillman . It would be great if we can include the update for MMF and completely remove all trace of av1c from v2 code base.

brhillman · 2021-08-27T00:18:26Z

Okay great, I'll work on that tomorrow @wlin7 thanks again for putting this together!

(F1850, F1950, F2010, and F20TR). The corresponding compsets that use CICE are renamed as F1850-CICE, F1950-CICE, F2010-CICE, and F20TR-CICE, respecticely. The F####-CICE compsets also set to go with SROF while the MPASSI counterparts with MOSART as river component.

wlin7 · 2021-09-02T14:34:15Z

With renamed F-compsets following the commit of 864ad0c, it has been verified that the new F2010 and F20TR are BFB with previous F2010-MPASSI and F20TR-MPASSI, respectively. Similarly, the new F2010-CICE and F20TR-CICE are BFB with previous F2010 and F20TR, respectively.

A_WCYCL1850-DIB-ISMF_CMIP6 has been replaced by CYRO1850. A_WCYCL1850-DIB_CMIP6 is not needed in v2.

jonbob · 2021-09-02T20:24:08Z

@wlin7 - the A_BGWCYCL1850 compset was just added and should not be removed. It's a cryo compset and not BGC (though it's hard to tell from the naming). Can it be added back, but without the A_ prefix?

wlin7 · 2021-09-03T03:43:27Z

@wlin7 - the A_BGWCYCL1850 compset was just added and should not be removed. It's a cryo compset and not BGC

Thanks, @jonbob . I didn't realize it was recently added. I removed it because it was using AV1C. I am going to change it to EAM%CMIP6 as you suggested.

jonbob

Approved based on visual inspection

jonbob · 2021-08-24T22:00:21Z

cime_config/allactive/config_compsets.xml

-  <alias>A_WCYCL1850-DIB-ISMF</alias>
-  <lname>1850_EAM%AV1C-L_ELM%SPBC_MPASSI%DIB_MPASO%IBISMF_MOSART_SGLC_SWAV</lname>
-</compset>
-
 <compset>
  <alias>A_WCYCL1850-DIB_CMIP6</alias>


The cryo people would like to keep the A_WCYCL1850-DIB-ISMF_CMIP6, though it's also OK to drop the "A_" prefix

) Update and clean-up compsets for use in test suites Remove obsolete eam and allactive compsets and the associated eam use_case files, and update the compsets used in nightly test suites. All compsets referring to AV1C or v0atm are removed. F20TRC5-CMIP6 and F2010SC5-CMIP6 are no more needed (previously retained for ongoing testings). Remainging use case files referring to av1c are removed, except for two used by MMF tests. This PR also makes all standard F-cases use prescribed data ice mode of MPASSI. This includes F1850, F1950, F2010, F20TR. The corresponding compsets that use CICE are renamed as F1850-CICE, F1950-CICE, F2010-CICE, and F20TR-CICE, respecticely. The F####-CICE compsets set to go with SROF while the MPASSI counterparts with MOSART as river component. The following compsets are replaced in nightly test suites. . FC5AV1C-L with F2010 . A_WCYCL1850 and A_WCYCL1850S_CMIP6 with WCYCL1850 . FC5AV1C-H01A with F2010 as there is no v2 compset just for hires . A_WCYCL1950S_CMIP6_HR and A_WCYCL2000_H01AS with WCYCL1950 as a placeholder [non-BFB] Virtually for all tests involving the compsets above. Tests using the old compsets would appear missing while new tests would appear correspondingly.

wlin7 · 2021-09-07T19:01:46Z

Merged to next.
Conflict in cime_config/allactive/config_compsets.xml due to new CRYO1950. Conflicts in components/eam/cime_config/config_compsets.xml due to naming changes among F###, F###-CICE and F###-MPASSI (some of which introduced in #4483).

…t (PR #4488)" This reverts commit 320c6c5, reversing changes made to 974108b.

wlin7 · 2021-09-08T14:38:19Z

Reverted due to missing input files for F2010 to go with test grids (mostly ne4_ne4).

…into next (PR #4488)"" This reverts commit c6b6220.

wlin7 · 2021-09-13T00:12:14Z

Reverted the previous revert of this merging this PR, then remerged the branch to next.
No conflicts on remerging for cime_config/allactive/config_compsets.xml or components/eam/cime_config/config_compsets.xml

conflicts in components/eam/cime_config/config_component.xml due to changes in compset matching pattern (additional character as in '[1XO]') for the use cases scam_arm97_MMF-1mom and scam_rico_MMF-1mom. Resolved to what are in the current master.

wlin7 · 2021-09-13T12:03:11Z

Tests return:
Chrysalis DIFFs expected and FAILs are accounted for.

ERS_Ln9.ne4pg2_ne4pg2.F-MMF1-RCEMIP.chrysalis_intel due to having use_htefrz_classnuc = .true.
SMS_D_Ln5.conusx4v1_conusx4v1.F2010.chrysalis_intel due to not having fsurdat for 2010, and likely no suitable grid for MPASSI

Compy and Mappy FAIL due to failing to download inputdata/ice/mpas-cice/oQU240/mpas-cice.graph.info.151209.part.48. The file is available on inputdata server. Manually copied to compy's inputdata path.

ERP_Ld3.ne4_oQU240.F2010.compy_pgi
ERP_Ld3.ne4_oQU240.F2010.mappy_gnu

chrysalis atm_nbfb_next FAILs: mpas-cice.graph.info.151209.part.2880

wlin7 · 2021-09-13T17:30:27Z

Merged to master.
Requests submitted separately for those requiring blessing (MMF tests), and those needing new baselines.

F2010 and F20TR wcprod tests on anvil failed. These tests passed on chrysalis and compy.

wlin7 · 2021-09-14T12:25:50Z

ERS_Ln9.ne4pg2_ne4pg2.F-MMF1-RCEMIP.chrysalis_intel passed (not a DIFF) after fixing compset matching for the use_case (via #4535). conusx4v1 test becomes a DIFF after #4535.

ERP_Ld3.ne4_oQU240.F2010 returned an expected diff (new test) on compy, but continued to fail on mappy for a different and unclear reason. The mappy test appears to have completed the run and trying to compare with non-existent baseline.

wlin7 · 2021-09-14T12:53:26Z

wcprod tests with F2010 and F20TR continued to fail on anvil (while no problem on other machines). Follow #4538 for more details and fixes.

chrysalis atm_nbfb_next: after having graph2880 file, still has two FAILs below -- MVK_PS test passed but still needs to be blessed (request submitted).

PGN_P1x1.ne4_oQU240.F2010.chrysalis_intel
TSC.ne4_oQU240.F2010.chrysalis_intel

TSC.ne4_oQU240.F2010 needs a new graph file: mpas-cice.graph.info.151209.part.1152.
PGN_P1x1.ne4_oQU240 failure due to an inconsistent finidat (to be fixed).

wlin7 · 2021-09-15T15:07:58Z

ERP_Ld3.ne4_oQU240.F2010.mappy_gnu was blessed anyway on Sept. 14. But the test continues being reported as a FAIL today, despite the comparison of cpl.hi with baseline passed and no apparent error or failure in the test log.

rljacob · 2021-09-15T15:50:13Z

@jgfouca can you figure out why ERP_Ld3.ne4_oQU240.F2010.mappy_gnu is flagged as "FAIL" ?

rljacob · 2021-09-15T16:03:14Z

This ERP test is PASS on: chrysalis, anvil, compy
FAIL on: mappy, cori-knl, cori-haswell
the fails on cori are without baselines so something in the test itself is triggering "FAIL". The only thing I see is memleak but I thought those were ignored.

wlin7 · 2021-09-17T16:43:33Z

Request submitted to bless anvil wcprod F2010 and F20TR tests, after the new default pelayout (#4541) fixed the runtime issue.

wlin7 · 2021-09-21T21:27:01Z

The tests below failed after this PR. Theta also did these tests.

SMS.ne120_oRRS18v3.WCYCL1950.cori-knl_intel.eam-cosplite
SMS.ne120_ne120.F2010.cori-knl_intel

wlin7 · 2021-09-28T17:26:17Z

After updated to latest cime master with ESMCI's #4092, rhe following atm_nbfb tests executed and returned DIFF, though aborted during baseline comparison. Request submitted to bless the diffs

PGN_P1x1.ne4_oQU240.F2010.chrysalis_intel
TSC.ne4_oQU240.F2010.chrysalis_intel

rljacob · 2021-09-28T17:47:24Z

I alread did that in https://acme-climate.atlassian.net/browse/SES-1814

…4551) Update V2 benchmarking grids and compsets (in part by #4488 too): low-res F: ne30_ne30.FC5AV1C-L -> ne30pg2_EC30to60E2r2.F2010 low-res G: T62_oEC60to30v3.GMPAS-IAF > T62_EC30to60E2r2.GMPAS-IAF low-res WC: ne30_oECv3.A_WCYCL1850S_CMIP6 -> ne30pg2_EC30to60E2r2.WCYCL1850 high-res F: ne120_ne120.FC5AV1C-H01A -> ne120pg2_r05_EC30to60E2r2.F2010 high-res WC: ne120_oRRS18v3.A_WCYCL1950S_CMIP6_HR -> ne120pg2_r0125_oRRS18to6v3.WCYCL1950 [BFB]

wlin7 added Atmosphere compset labels Aug 24, 2021

wlin7 added this to the v2.0 milestone Aug 24, 2021

wlin7 requested review from rljacob, jonbob and golaz August 24, 2021 20:40

wlin7 self-assigned this Aug 24, 2021

rljacob reviewed Aug 24, 2021

View reviewed changes

rljacob approved these changes Aug 24, 2021

View reviewed changes

wlin7 requested a review from brhillman August 24, 2021 20:52

jonbob reviewed Aug 24, 2021

View reviewed changes

wlin7 added 2 commits September 2, 2021 09:19

Remove compsets A_WCYCL1850-DIB_CMIP6 and A_WCYCL1850-DIB-ISMF_CMIP6

97c5292

A_WCYCL1850-DIB-ISMF_CMIP6 has been replaced by CYRO1850. A_WCYCL1850-DIB_CMIP6 is not needed in v2.

Renamed A_BG1850CN as BG1850CN

475fe28

Restored compset BGWCYCL1850 and replaced AV1C-L phys

3dc3f32

jonbob approved these changes Sep 3, 2021

View reviewed changes

wlin7 added a commit that referenced this pull request Sep 8, 2021

Revert "Merge branch 'wlin/atm/cleanup_v2compsets_for_tests' into nex…

c6b6220

…t (PR #4488)" This reverts commit 320c6c5, reversing changes made to 974108b.

brhillman and others added 5 commits September 10, 2021 12:56

Update MMF use case files for 2010 forcings

0f5c969

Update MMF F-cases to F2010

6dba9f4

Remove some unused use cases

66318c3

Remove cam4 use cases files

c06d4ce

Replaced F2010 ERP_Ln# tests with ERP_Ld3

7d6fa20

wlin7 added a commit that referenced this pull request Sep 12, 2021

Revert "Revert "Merge branch 'wlin/atm/cleanup_v2compsets_for_tests' …

c994286

…into next (PR #4488)"" This reverts commit c6b6220.

wlin7 merged commit 1a02264 into master Sep 13, 2021

wlin7 mentioned this pull request Sep 14, 2021

Reset grid and compset for F2010 tests with RRM #4535

Merged

This was referenced Sep 14, 2021

Update pe-layout for anvil's F2010 and F20TR wcprod tests #4538

Closed

atm_nbfb tests need reworked after switching to using new F2010 with mpassi #4540

Closed

wlin7 mentioned this pull request Sep 15, 2021

Change default layout for ne30pg2_oECv3 F-cases on anvil #4541

Merged

rljacob deleted the wlin/atm/cleanup_v2compsets_for_tests branch September 16, 2021 15:44

wlin7 mentioned this pull request Sep 16, 2021

Update PGN and TSC test scripts for E3SM ESMCI/cime#4092

Merged

This was referenced Sep 20, 2021

memory leak for ERP_Ld3.ne4_oQU240.F2010 #4546

Closed

Remove extra history write and reorg #4544

Merged

amametjanov mentioned this pull request Sep 22, 2021

Update V2 benchmarking grids and compsets #4551

Merged

wlin7 mentioned this pull request Sep 13, 2022

P3 cloud microphysics scheme #5168

Merged

Update and clean-up compsets for use in test suites #4488

Update and clean-up compsets for use in test suites #4488

Conversation

wlin7 commented Aug 24, 2021 • edited Loading

Choose a reason for hiding this comment

wlin7 Aug 24, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rljacob commented Aug 24, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

wlin7 commented Aug 24, 2021 • edited Loading

brhillman commented Aug 26, 2021

wlin7 commented Aug 27, 2021

brhillman commented Aug 27, 2021

wlin7 commented Sep 2, 2021

jonbob commented Sep 2, 2021

wlin7 commented Sep 3, 2021

jonbob left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

wlin7 commented Sep 7, 2021

wlin7 commented Sep 8, 2021

wlin7 commented Sep 13, 2021

wlin7 commented Sep 13, 2021 • edited Loading

wlin7 commented Sep 13, 2021

wlin7 commented Sep 14, 2021

wlin7 commented Sep 14, 2021 • edited Loading

wlin7 commented Sep 15, 2021 • edited Loading

rljacob commented Sep 15, 2021

rljacob commented Sep 15, 2021

wlin7 commented Sep 17, 2021

wlin7 commented Sep 21, 2021

wlin7 commented Sep 28, 2021

rljacob commented Sep 28, 2021

wlin7 commented Aug 24, 2021 •

edited

Loading

wlin7 Aug 24, 2021 •

edited

Loading

wlin7 commented Aug 24, 2021 •

edited

Loading

wlin7 commented Sep 13, 2021 •

edited

Loading

wlin7 commented Sep 14, 2021 •

edited

Loading

wlin7 commented Sep 15, 2021 •

edited

Loading